fix(FR/IT): normalise state.level field for region/department hierarchy

[dr5hn]

Summary

Customer-facing follow-up to #1349 / #1352. Cities are already correctly parented onto departments (FR) and provinces (IT) via the recently merged remap PRs (#1395, #1394, #1393, #1400, #1484), but the state rows themselves still carried inconsistent level values. This PR normalises the field so downstream consumers can reliably filter level=1 (region) vs level=2 (department/province).

Driven by bin/scripts/fixes/states_level_normalise.py (idempotent, only touches level, only on FR/IT).

Before / After

France (FR)

level	Before	After
null	29	0
1	0	29
2	95	95

29 rows null → 1: 13 metropolitan regions, 2 metropolitan collectivities with special status (Paris, Métropole de Lyon), 1 European collectivity (Alsace), 5 overseas regions, 5 overseas collectivities, 1 overseas collectivity with special status (Nouvelle-Calédonie), 1 overseas territory (TAAF), 1 dependency (Clipperton). 95 metropolitan departments unchanged at 2.

Italy (IT)

level	Before	After
null	1	0
1	115	20
2	4	106
3	6	0

103 rows updated. Final state: level=1 is region (15) + autonomous region (5) = 20. level=2 is province (80) + metropolitan city (14) + free municipal consortium (6) + decentralized regional entity (4) + autonomous province (2) = 106.

Validation

• Schema unchanged: only the level field was rewritten on 132 FR/IT rows; every other field byte-for-byte intact.
• Cross-reference: country_id ↔ country_code consistent for all 250 FR/IT states; no duplicate ids introduced.
• Idempotent: re-running the script reports zero changes.
• Parent-id integrity check ran post-normalisation. After the fix, every FR metropolitan department (level=2) has a parent_id resolving to a level=1 FR row, and every IT province / metropolitan city / free municipal consortium / decentralized regional entity / autonomous province (level=2) has a parent_id resolving to a level=1 IT row — with one pre-existing exception unrelated to this PR: 8 Tuscany provinces (Pisa, Pistoia, Prato, Siena, Livorno, Lucca, Massa and Carrara, Grosseto, ids 1685/1687/1700/1734/1745/1749/1759/1787) carry parent_id=1764 ("Udine", a Friuli decentralized regional entity) instead of 1664 ("Tuscany"). This is a stale parent_id from feat(IT): remap cities to metropolitan cities and provinces (#1349) #1395 and is orthogonal to level normalisation; flagging here for a follow-up cleanup.

Notes

• Refs [Bug]: Italy seems to be totally wrong #1349 and [Bug]: France data — missing cities and regions misclassified #1352. Does not close either.
• No id / created_at / updated_at / flag fields were added or modified.
• Non-FR/IT countries: untouched.

[Copilot]

Pull request overview

Normalises state.level for France and Italy so consumers can reliably distinguish region-tier (level=1) vs department/province-tier (level=2) states, aligning state metadata with the already-correct city parenting from prior remap PRs.

Changes:

• Rewrites inconsistent/missing level values on FR/IT state records in contributions/states/states.json to a consistent 1 (region tier) or 2 (department/province tier).
• Adds an idempotent fix script (bin/scripts/fixes/states_level_normalise.py) to apply and validate the normalisation (including a parent-id integrity diagnostic).

Reviewed changes

Copilot reviewed 1 out of 2 changed files in this pull request and generated no comments.

File	Description
contributions/states/states.json	Normalises FR/IT level values (regions/collectivities → 1, departments/provinces → 2).
bin/scripts/fixes/states_level_normalise.py	Adds a repeatable script to compute/apply the level normalisation and report unmapped types / parent-id issues.